Hannu Krosing <hannu@tm.ee> writes:
> There a bug report that allowed tuplicate ids in an uniqe field when
> SELECT FOR UPDATE was used. Could this be your case ?
> [snip]
> IIRC the fix was also provided, so it could be fixed in current CVS (the
> above is from 7.0.2, worked the same in 6.5.3)
It does seem to be fixed in current CVS:
regression=# create table test(i int primary key);
NOTICE: CREATE TABLE/PRIMARY KEY will create implicit index 'test_pkey' for table 'test'
CREATE
regression=# insert into test values(1);
INSERT 145215 1
regression=# begin;
BEGIN
regression=# select * from test for update;i
---1
(1 row)
regression=# insert into test values(1);
ERROR: Cannot insert a duplicate key into unique index test_pkey
regression=#
regards, tom lane